Master Thesis

Designing Biometric Identification Interface with Palm Vein Structure on Embedded System, 2015




Palm vein images are agreeable as a model which may give appropriate patterns to biometric classification. Palm veins that predecessor in point of uniqueness, are a biometric specification, studies work around methods of imaging, preprocessing procedures, methods of taking feature vectors and classification methods. Those studies go through with one or many of priorities which accuracy, processing speed, portability, and security. In this work, vein image capture system hardware had designed and prototype was produced, then with this system, genuine database was created with combining age and gender of test subjects. On this database, popular preprocessing methods, feature vectors taking methods and classification methods were tested and results are compared with speed vs accuracy. All the subsections which used in hardware and software defined as modules. Python has been used as software module programming on SBC (Single Board Computer) with Linux kernel. As a result, appropriate system has been built for both secured software and hardware interfaces with accuracy of biometric classification results of palm vein structure.

Journal Articles

Palm Vein Authentication and Verification System, 2015

Beyin Bilgisayar Arayüzü için Dvm Makine Öğrenme Yöntemi Kullanilarak Eeg Verilerinden Sağ Ve Sol El Hareket Düşüncelerinin Tespiti, 2017




Brain-computer interface (BCI) is a direct communication channel between computer and human brain. BCI can be used as a part of prothesis and communication technologies, allowing severely disabled people to communicate with machines directly via BCI. In this study, a BCI system is designed that can detect mental imagery of right and left hand movements from EEG data, by using a modified portable Epoc Emotiv EEG headset and the Supported Vector Machines which one of machine learning method. Designed BCI system is found to have approximately 80% accuracy per single imagery event, for separating right and left hand movement imageries. Thus proposed BCI system can offer 90-95% accuracy if used with two imagery events, sufficient for enabling the BCI control of external devices.

Conference Papers

Mersin ili Yenisehir İlcesi ve Bagli Bolgelerinde Elektromanyetik Kirlilik Olcum Calismalari ve Haritalari,EMANET ,2015






Mersin Yenişehir ilçesi ve bağlı bölgelerinde baz istasyonlarının ve diğer elektromanyetik alan kaynaklarının yaydığı elektrik alan şiddetlerinin ortalama ölçümleri yapılmış ve yerleşkenin 900 ve 1800 MHz frekans bazlı şiddetleri içeren elektromanyetik alan yoğunluk haritaları çıkarılmıştır. Yerleşke dâhilinde 1500’ün üzerinde noktada her iki frekans bileşeni için anlık ve ortalama elektromanyetik alan yoğunluğu ölçülmüştür. Elde edilen sonuçlarla, elektromanyetik alan yoğunlukları ayrı ayrı haritalanmış ve sonuçlar değerlendirilmiştir.